Trimethylamine Gas Sensor Based on Electrospun In₂O₃ Nanowires with Different Grain Sizes for Fish Freshness Monitoring

Abstract

Seafood, especially marine fish, is highly prone to spoilage during processing, transportation, and storage. It releases pungent trimethylamine (TMA) gas, which severely affects food quality and safety. Metal–oxide–semiconductor (MOS) gas sensors for TMA detection offer a rapid, convenient, and accurate method for assessing fish freshness. Indium oxide (In₂O₃) has shown potential as an effective sensing material for the detection of TMA. In this work, one-dimensional In₂O₃ nanowires with different grain sizes and levels of crystallinity were synthetized using the electrospinning technique and underwent different thermal calcination processes. Gas-sensing tests showed that the In₂O₃−3 °C/min−500 °C gas sensor exhibited an outstanding performance, including a high response (R_a/R_g = 47.0) to 100 ppm TMA, a short response time (6 s), a low limit of detection (LOD, 0.0392 ppm), and an excellent long-term stability. Furthermore, the sensor showed promising experimental results in monitoring the freshness of Larimichthys crocea (L. crocea). By analyzing the relationship between the grain size and crystallinity of the In₂O₃ samples, a mechanism for the enhanced gas-sensing performance was proposed. This work provides a novel strategy for designing and fabricating gas sensors for TMA detection and highlights their potential for broad applications in real-time fish freshness monitoring.

1. Introduction

With the rapid development of the economy and improvements in living standards, people’s dietary habits and food varieties have become increasingly diversified. A rich variety of marine and aquatic products has entered public consumption, among which marine fish are particularly favored by consumers due to their high contents of protein, amino acids, vitamins, and minerals [1]. Meanwhile, consumers are placing higher demands on food freshness and safety. However, marine fish are highly susceptible to spoilage during processing, transportation, and storage, which significantly compromises their quality and poses risks to consumer health [2,3]. Therefore, there is an urgent need for rapid and effective methods to detect and evaluate the freshness of marine fish [4,5].

Generally, fish undergo physical and chemical changes during the spoilage process, which involve changes in appearance, color, odor, and microbial colonies. Therefore, the freshness of fish can be assessed through various techniques, including sensory characteristic analysis (color and texture) [6], physical and chemical experimental analysis (pH level, K value, and total volatile basic nitrogen (TVB-N)) [7,8], and gas analysis (volatile organic compounds, VOCs) [9,10]. Among these, sensory characteristic analysis is a traditional detection method, with the quality index method (QIM) being the most widely used [11,12]. This method evaluates the color, taste, and tissue morphology of the fish surface and summarizes the results to ultimately determine the quality level. Physical and chemical experimental analyses linked to chemical composition changes in fish are essential and indispensable methods for evaluating fish freshness. Common physical and chemical experimental analyses include moisture content analysis [13,14], TVB-N determination [15,16], protein degradation assessment [17,18], lipid oxidation evaluation [19,20], and K-value measurement based on adenosine triphosphate (ATP) degradation [21,22]. These indicators collectively provide a comprehensive framework for assessing freshness in fish products. Although the previously mentioned freshness detection techniques are relatively accurate, they have limitations such as being destructive, expensive, and time-consuming [23]. Therefore, there is a need for a convenient, rapid, and accurate method to assess fish freshness.

It is important to mention that trimethylamine oxide (TMAO), an abundant compound in marine fish, is converted into TMA under the action of bacterial enzymes, which is the primary cause of fishy odors [24]. Therefore, the measurement of TMA in seafood products is an indicator for assessing the freshness of fish (fresh: 0–1 mg/100 g, < 10 ppm; beginning to spoil: 1–5 mg/100 g, 10–50 ppm; and severely spoiled: > 6 mg/100 g, > 60 ppm) [5,25]. In recent years, various methods for detecting TMA have been extensively developed and applied, such as high-performance liquid chromatography (HPLC), gas chromatography–mass spectrometry (GC-MS), and ion mobility spectrometry [26,27]. However, these detection methods usually require expensive precision instruments and complex operating procedures, making them unsuitable for on-site assessments of fish freshness. The MOS gas sensor is a novel method for detecting TMA, which can simply analyze the concentration of the gas with a high sensitivity and low cost [28,29].

MOSs have attracted widespread attention due to their low cost, good stability, and excellent electrochemical performance, and are widely used in fields such as gas sensors [30] and electrode materials [31]. Researchers have developed various MOS gas sensors for TMA detection, including materials such as WO₃, ZnO, MoO₃, α-Fe₂O₃, SnO₂, and In₂O₃. Shen et al. [32] designed a TMA gas sensor based on oxygen-vacancy-rich and porous α-MoO₃ nanosheets, which demonstrated a good linear response to 0.02–50 ppm TMA at 133 °C and an LOD of as low as 20 ppb. At its optimal operating temperature, the sensor exhibited a response value of 51.47 to 10 ppm TMA, with response/recovery times of 15 s and 300 s, respectively, highlighting its potential for practical applications. Liu et al. [33] synthesized α-Fe₂O₃ nanoparticles (~500 nm in diameter) via a solvothermal method and tested their gas-sensing performance as thick-film sensors. At 250 °C, the α-Fe₂O₃ nanoparticles showed an excellent selectivity and high sensitivity to TMA, with a minimum detectable concentration of 1 ppm. The sensor achieved a response value of 27.8 to 100 ppm TMA within 1 min, with a recovery period of less than 4 min. Wang et al. [34] prepared ordered mesoporous WO₃ and Au-loaded WO₃ (Au−WO₃) sensors using KIT-6 as a template. The Au−WO₃ sensor exhibited an exceptional sensing performance at 268 °C, including a high response (R_a/R_g = 41.56) to 100 ppm TMA, an ultrafast response speed (1 s), and an excellent selectivity, though its recovery time was 323 s.

According to the above review, MOS gas sensors have advantages such as portability, cost-effectiveness, and a high sensitivity, making them a promising method for assessing the freshness of fish. However, MOS gas sensors still have issues such as a low sensitivity to TMA and long response/recovery times during the detection process. In₂O₃ material is a typical n-type MOS material. Compared with other MOSs, it has the advantages of a good thermal stability and low resistivity. It also has the advantages of safety, non-toxicity, low cost, and a good gas sensitivity to various oxidizing and reducing gases. These properties have enabled its widespread application in optoelectronics, gas sensing, and catalysis [35,36,37]. The short response/recovery time and long-term stability of In₂O₃ make it a leading candidate for advancing TMA-sensing technologies in practical applications. Researchers exploring In₂O₃-based gas sensors have identified significant improvements. For instance, Lu et al. [38] synthesized flower-shaped In₂O₃ nanostructures using a hydrothermal method. The sensor achieved its maximum TMA sensitivity at an optimal operating temperature of 340 °C, with short response/recovery times of 5 s and 10 s. These results indicate that the sensor is highly sensitive to TMA and can effectively be utilized to assess the freshness of fish. Li et al. [39] developed a novel flexible TMA gas sensor using In₂O₃ nanofibers. At 80 °C, the sensor exhibited a high response (3.8 to 1 ppm TMA), short response/recovery time (6 s and 10 s), and excellent reliability and flexibility, further confirming the suitability of In₂O₃ for high-performance TMA detection.

In recent years, researchers have investigated various synthesis methods to create In₂O₃-based nanomaterials with different microstructures to improve gas sensor performance. Wang et al. [40] prepared one-dimensional In₂O₃ nanofibers via electrospinning, further increasing the specific surface area. Gas-sensing tests revealed that these In₂O₃ nanowires exhibited a high sensitivity, selectivity, and rapid response/recovery speed when exposed to ethanol vapor. Santucci et al. [41] synthetized In₂O₃ thin films through high-vacuum thermal evaporation and sol–gel techniques, showing their superior gas-sensing response to NO₂ at 250 °C. Building on this, Arbiol et al. [42] investigated the impact of structural parameters (e.g., film thickness and grain size) on gas-sensing performance, concluding that reducing film thickness and grain size significantly enhanced the sensitivity of In₂O₃-based gas sensors. These studies highlight the critical role of tailored nanostructures in advancing the application of In₂O₃ for gas detection technologies.

In this study, a simple electrospinning method was employed to synthesize one-dimensional In₂O₃ nanowires. The microstructure of the In₂O₃ nanowires was adjusted by changing the heating rates and annealing temperatures during the thermal treatment process, which affected their surface and grain sizes. Among the samples compared, the one heated to 500 °C at a rate of 3 °C/min exhibited a higher response (47.0), faster response speed (6 s), and lower LOD (0.0392 ppm). Based on an analysis of characterization results and gas sensitivity tests, the observed enhancement in TMA gas-sensing performance can be attributed to the alterations in the surface and grain size of the In₂O₃ nanowires induced by the changes in heating rate and annealing temperature. This work provides a novel approach for the development of high-performance TMA gas sensors.

2. Experimental Section

This section provides a detailed introduction to the materials required for the preparation of In₂O₃ nanowires and outlines the process for their synthesis. It also describes the techniques employed to investigate the morphology and composition of the gas-sensitive materials, as well as to test their gas-sensing characters.

2.1. Materials

Indium nitrate hydrate (In(NO₃)₃·4.5H₂O) and N,N−dimethylformamide (DMF) were purchased from Sinopharm Chemical Reagent Co., Ltd., Beijing, China. PVP (Mw = 1,300,000) was purchased from Aladdin Biochemical Technology Co., Ltd., Shanghai, China.

2.2. Synthesis of In₂O₃ Nanowires

A total of five types of In₂O₃ nanowires were prepared as sensing materials in this work, all of which were synthetized by the electrospinning method. The synthesis process of each sample is described as follows.

First, in the preparation of the precursor solution for electrospinning, 0.3 g of In(NO₃)₃·4.5H₂O was added to 3 mL of DMF and stirred thoroughly until a transparent solution was formed. Then, 0.8 g of PVP was gradually introduced into the solution under continuous stirring. After 10 h of stirring, the PVP was completely dissolved, resulting in a colorless, viscous, and transparent precursor solution, which was used for the subsequent synthesis of In₂O₃ nanowires. The obtained transparent precursor solution was then loaded into a 10 mL disposable syringe equipped with a specialized electrospinning needle. Following that, the syringe was securely mounted onto the electrospinning machine. The needle was connected to a DC power supply set at 11 kV, and the feed rate of the syringe was adjusted to 0.3 mL/h. The distance between the needle tip and the drum collector was fixed at 14 cm. Under controlled environmental conditions (humidity: 30–40% and temperature: 30 °C), electrospinning was carried out for 3–4 h, and a uniform white film was covered on the tin foil. Finally, the film was carefully peeled off using tweezers and transferred to a square ceramic boat for further processing.

Finally, the rectangular ceramic boat was placed into a muffle furnace for calcination. This study primarily investigated the effects of different calcination conditions on In₂O₃ materials. During this process, various In₂O₃ nanowires were obtained by altering the heating rate and annealing temperature. For instance, the muffle furnace was set to heat at a rate of 3 °C/min, reaching a temperature of 500 °C, and calcined for 2 h to obtain a pale yellow In₂O₃ powder sample, labeled as In₂O₃−3 °C/min−500 °C. Following this method, four other samples were prepared and labeled as In₂O₃−1 °C/min−500 °C, In₂O₃−5 °C/min−500 °C, In₂O₃−3 °C/min−400 °C, and In₂O₃−3 °C/min−600 °C, respectively.

2.3. Characterization

X-ray powder diffraction (XRD) data were obtained with the D2 PHASER X-ray diffractometer (AXS, Bruker, Karlsruhe, Germany) with high-intensity Cu K radiation in the range of 9–90°. Field emission scanning electron microscopy (FESEM) images were obtained on a ZEISS GX1669 microscope (Carl Zeiss AG, Oberkochen, Germany). X-ray photoelectron spectroscopy (XPS) data were recorded on a Escalab 250Xi system (Thermo Fisher Scientific, Waltham, MA, USA).

2.4. Gas Sensor Fabrication and Test

The fabrication and measurement of the gas sensor in this work are consistent with our previous studies [43,44]. A tubular device was employed to manufacture the gas sensor, the main structure of which was an alumina ceramic tube (4 mm in length, 1.2 mm in outer diameter, and 0.8 mm in inner diameter). This tube was symmetrically wound with a pair of annular Au electrodes on both sides. A Pt wire was drawn from each side of every Au electrode. Initially, approximately 3–4 mg of In₂O₃ sample was mixed with deionized water at a 3:1 mass ratio and ground into a paste using an agate mortar. Subsequently, the prepared paste was evenly applied many times in a small amount onto the Al₂O₃ ceramic tube, and the tube with sensing material on the surface was baked under infrared light for 20 min to form a uniform film on its surface. Following this, the ceramic tube coated with the sensing material film was sintered at 200 °C for 2 h to eliminate all traces of moisture and to further solidify the sensing material on the ceramic tube surface. After cooling to room temperature, a Ni−Cr alloy coil was placed into the cavity of the ceramic tube to serve as the heat source for the sensor. Finally, the ceramic tube and alloy coil were soldered to the sensor socket accordingly, completing the gas sensor fabrication process.

The gas-sensing performance of the fabricated gas sensors was evaluated in a laboratory environment (30% RH, 25 °C) using a commercial gas-sensing test platform (Weisheng Electronics Technology Co., Ltd., model WS−30A, Zhengzhou, China, test chamber volume: 18 L). The amount of liquid VOCs in the sample was calculated according to Equation (1), where C represents the concentration of the analyte (ppm), φ denotes the purity of the liquid (%), ρ is the density of the liquid (g/cm³), V₁ is the volume of the liquid (μL), V₂ is the volume of the test chamber (L), and M is the molecular weight of the liquid (g/mol) [45].

$$\mathrm{C}={\displaystyle \frac{22.4\times \phi \times \mathsf{\rho }\times {\mathrm{V}}_1}{\mathrm{M}\times {\mathrm{V}}_2}}\times 1000$$

(1)

First, the electrical current flowing through the heating wire was adjusted to ensure that the device operated at the pre-determined temperature. Next, a micro-syringe was employed to precisely extract the calculated volume of TMA liquid, which was then injected into the evaporator. After sealing the lid and turning on the fan, a TMA gas atmosphere with a specific concentration was formed once the liquid had completely evaporated and uniformly dispersed. At the same time, the testing software automatically recorded the resistance changes of the device. Finally, after the resistance stabilized for a few minutes, the seal was opened to release any remaining gas. Throughout the test process, the ambient temperature and relative humidity were maintained at approximately 25 °C and 30%, respectively.

The response of the sensor is defined as S = R_a/R_g (R_a and R_g are the resistances of the sensor when it is stable in air and TMA, respectively) and the response time (τ_res.) and recovery time (τ_rec.) are defined as the time required for the resistance of the sensor to change by 90% in TMA and air, respectively.

As shown in Figure S1, the experimental device for detecting the spoilage of L. crocea was a manually built test system. Before the test began, the sensor was first connected to a multimeter (Fluke 8846a) to monitor its real-time resistance changes. The multimeter was connected to a computer to display and record the response/recovery curve of the sensor. The heating wire of the sensor was connected to the power supply (GPD-4303S), and the working temperature of the sensor was adjusted by controlling the current passing through the heating wire. In one test, the R_a of the sensor was first recorded, and the sensor was then placed in a sealed container storing L. crocea when R_a was stabilized for a period of time and its resistance changes had been recorded. After the test was completed, the sensor was removed from the container and it began to recover. The above steps were repeated and the TMA concentration of L. crocea was monitored every 24 h.

3. Results and Discussion

3.1. Structural and Morphological Features

To confirm the material composition and crystallinity of the synthesized In₂O₃ samples, XRD analysis is conducted on the samples. Figure 1 displays the XRD diffraction patterns of the samples at different heating rates and annealing temperatures. The diffraction peaks of all samples are generally consistent with those of In₂O₃ (JCPDS Card No. 89−4595), and no diffraction peaks related to other impurities are detected. This indicates that the synthesized samples are free of any impurities and consist of pure In₂O₃.

It has been reported that differences in heating rate and annealing temperature can affect the crystallinity and grain size of samples [46,47]. As shown in Figure 1, when the heating rate remains constant, the diffraction peak intensity of the samples gradually increases with a higher annealing temperature. When the annealing temperature remains constant, the diffraction peak intensity of the samples initially rises and then falls with an increase in the heating rate. Additionally, the average grain size of each sample is calculated using the Scherrer Equation, $\mathrm{D}=0.9\lambda /\beta \cos \theta$, where λ is the X-ray wavelength of CuKα (1.5418 Å), θ is the Bragg angle, and β is the full width at the half maximum of a peak [48]. These results are summarized in

Table 1. Summary of the grain size in In₂O₃ nanowire samples with different annealing temperatures and heating rates.

Sample	Grain Size (nm) a
In2O3−1 °C/min− 500 °C	15.30
In2O3−3 °C/min− 500 °C	9.06
In2O3−5 °C/min− 500 °C	12.67
In2O3−3 °C/min− 400 °C	7.66
In2O3−3 °C/min− 600 °C	12.80

^a Results calculated from XRD analysis.

, which shows that with a constant heating rate, the grain size increases with an increase in the annealing temperature, and when the annealing temperature remains constant, the grain size initially increases and then decreases as the heating rate increases.

The shape and structure of a sample can significantly influence its gas-sensing properties. Therefore, the microscopic structure of the samples is characterized using SEM. Figure 2 displays the SEM images of the five In₂O₃ nanowires under different heat treatment conditions. Overall, the In₂O₃ nanowires are radially uniform with axial lengths exceeding 10 μm, although there are slight variations in radial size among different nanowires. The formation of these In₂O₃ nanowires is mainly caused by the decomposition of polymers and the shrinkage of fibers during the calcination process. The precursor solution generates nanowires containing In³⁺/PVP through electrospinning. Under high-temperature calcination, PVP gradually decomposes and In³⁺ reacts with oxygen to crystallize and aggregate into In₂O₃ nanoparticles arranged along the fiber axis. Therefore, the sample ultimately forms a nanowire structure connected by In₂O₃ nanoparticles. As illustrated in Figure 2a–i, the In₂O₃ nanowires are subjected to sintering at different heating rates while being annealed at a temperature of 500 °C. When the heating rate is slow (1 °C/min), the extended sintering time allows the grains enough time to absorb nearby unstable phases, leading to significant grain coarsening and a gradual increase in grain size. Therefore, the nanowire surfaces are rough and the structure is dense. As the heating rate increases (3 °C/min), PVP begins to gradually decompose and metal ions in the nanowires form smaller nanocrystals, resulting in a rougher nanowire surface with uniformly distributed grains. When the heating rate is further increased (5 °C/min), PVP reaches its decomposition temperature more rapidly, reducing the template’s constraint on fiber morphology and grain size, which leads to nanowires with larger grain sizes and denser surfaces.

As shown in Figure 2d–f, j–o, the samples are sintered at different annealing temperatures with a heating rate of 3 °C/min to obtain In₂O₃ nanowires. When the annealing temperature is low (400 °C), the surface of the nanowires is smooth, with only a few grains present. Combined with XRD analysis, it is understood that, at 400 °C, the sintering process primarily involves the thermal decomposition of PVP, with only a small amount of In₂O₃ grains forming. This results in a lower crystallinity and smaller grain sizes in the nanowires. As the annealing temperature increases (500 °C), the crystallinity of the nanowires improves, the grain size further increases, and the surface becomes rougher. When the annealing temperature is further increased (600 °C), the nanowire surface appears denser with larger grain sizes. This demonstrates that different heat treatment methods have an impact on the microstructure of In₂O₃ nanowires. Structurally, the In₂O₃−3 °C/min−500 °C nanowires have a rougher surface and smaller grain sizes, which provide greater advantages and potential in gas-sensing applications. In addition, Figure S2 shows the surface morphology of the sensing device coated with In₂O₃−3 °C/min−500 °C nanowires after multiple gas-sensing tests. In Figure S2a, In₂O₃ nanowires uniformly cover the outer surface of the ceramic tube. Figure S2b–d show enlarged views of the In₂O₃ nanowires. It is noticed that there are many fractures among them, which may be due to external pressure on the long nanowires during the grinding process. However, most nanowires still maintain their one-dimensional structure and rough surface, demonstrating an excellent structural stability which is free from the influence of gas-sensing tests.

To delve deeper into the relationship between the internal characteristics of the materials and their gas-sensing properties, XPS is utilized to characterize the chemical states of the surface elements in the synthesized In₂O₃ nanowires. All peak positions are calibrated based on the standard C 1s peak (284.8 eV). Figure 3a displays the full-scan XPS spectra of the five samples, showing that the peaks and their positions are essentially the same across the samples, and all orbital peaks can be clearly identified. As shown in Figure 3b, two separate peaks are observed in the In 3d spectrum, corresponding to In3d_3/2 and In3d_5/2. Although the peak positions of each set of In 3d peaks vary due to differences in the internal chemical environments of the substances, the splitting energy between the two sets of peaks remains constant at 7.6 eV, which is consistent with the measured In 3d splitting energy of the samples.

Generally, the state of oxygen in nanowires affects the material’s ability to adsorb and ionize oxygen, thereby influencing gas detection performance. To analyze the state of oxygen in the In₂O₃ nanowires, the O 1s core level spectra of the samples are divided into three Gaussian components. As shown in Figure 4a–e, the rightmost peak (red) is identified as lattice oxygen (O_L), which is a fundamental part of the crystal. Generally, O_L oxygen ions within the lattice are very stable and do not contribute to the gas-sensing properties of the material. The middle peak (purple) and the leftmost peak (green) are referred to as defect oxygen (O_V) and adsorbed oxygen (O_C), respectively. It is generally believed that the more defects, active sites, and surface adsorbed oxygen there are, the more favorable this is for gas adsorption reactions on the surface of the gas-sensing material, leading to higher response values [49,50]. From the summary in

Table 2. Holistic fitting results of O 1s XPS spectra of five as-synthesized samples.

Sample	OL (%) a	OV (%) a	OC (%) a
In2O3−1 °C/min−500 °C	64.7	20.5	14.8
In2O3−3 °C/min−500 °C	48.4	31.6	20.0
In2O3−5 °C/min−500 °C	62.0	19.7	18.3
In2O3−3 °C/min−400 °C	54.0	27.1	18.9
In2O3−3 °C/min−600 °C	61.1	19.8	19.1

^a Results from XPS analysis.

, it is clear that the beneficial components for gas-sensing performance (O_V and O_C) in the In₂O₃−3 °C/min−500 °C nanowires account for 51.6% of all oxygen species, which enhances the potential of In₂O₃−3 °C/min−500 °C nanowires as high-performance gas-sensing materials.

3.2. Gas-Sensing Properties

Generally, the sensing characteristics of MOS gas sensors are influenced by the operating temperature of the sensor. As the temperature increases, the gas molecules overcome the activation energy barrier of the surface reaction and react with the oxygen species on the surface of the material, and the sensor response increases. When the optimal operating temperature is reached, the gas resolution rate is greater than the adsorption rate, and the response of the gas sensor decreases. It is essential to first determine the optimal operating temperature for the sensor. Figure 5 illustrates the dynamic response of the sensor to 100 ppm TMA at different temperatures ranging from 200 °C to 300 °C. Overall, the responses of all five sensors initially increase and then decrease, which is consistent with the characteristics of MOS-type gas sensors. Evidently, the In₂O₃−3 °C/min−500 °C gas sensor achieves a response value of 47 at an operating temperature of 240 °C, which is higher than the responses of the other four sensors. This phenomenon indicates that a smaller grain size contributes to an improved response while reducing the optimal operating temperature of the sensor, which is generally consistent with the previous assumptions. The In₂O₃−3 °C/min−500 °C gas sensor exhibits a notably superior sensing performance for TMA.

Response/recovery characteristics are also crucial metrics for evaluating the performance of gas sensors, particularly in real-world applications where rapid detection is essential. Figure 6a–e show the response and recovery curves of the five different sensors when exposed to 100 ppm TMA at their optimal operating temperatures. Overall, the sensors exhibit a relatively fast response and recovery speed. Among them, the In₂O₃−3 °C/min−500 °C and In₂O₃−3 °C/min−400 °C gas sensors have the shortest response times, reaching 90% of the signal change within just 6 s. This rapid response can be attributed to the sensors’ optimized surface morphology and enhanced active sites, which promote the rapid adsorption of gases. The response time is the same for both sensors, but they have slightly different recovery times: the In₂O₃−3 °C/min−500 °C gas sensor recovers in 37 s, while the In₂O₃−3 °C/min−400 °C gas sensor takes 39 s. This difference suggests that a higher annealing temperature may promote more efficient desorption due to increased crystallinity and reduced defect-related capture sites. The test results indicate that different heat treatment methods can directly influence the adsorption and desorption rates of the sensing material to gas. These findings reveal the importance of precise heat treatment in improving the sensitivity of gas sensors.

Figure 7a–e illustrate the response/recovery characteristics of the sensors with different concentrations of TMA at their optimal operating temperatures. The responses of the five different sensors increase with a rising concentration of TMA, indicating that the sensors possess a good recognition capability for different concentrations of TMA. Among them, the five different sensors still exhibit a certain response to low concentrations of TMA (1 ppm). For instance, the In₂O₃−3 °C/min−500 °C sensor can achieve a response value of 2.64 to 1 ppm of TMA. Figure 7f displays the linear fitting of the response−concentration curves for the five sensors. According to the slope of the fitted curve, the In₂O₃−3 °C/min−500 °C gas sensor exhibits a better sensitivity. Additionally, the theoretical LOD of the sensor can be calculated using the following equation: RMS_noise (ppm⁻¹) = (R²_s/(N − 1))^1/2, $\mathrm{LOD}=\left(3\times {\mathrm{RMS}}_{\mathrm{noise}}\right)/\mathrm{Slope}$, where ${\mathrm{R}}_{\mathrm{s}}^2$ and ${\mathrm{RMS}}_{\mathrm{noise}}$ are the quadratic sum of the regular residual and the root mean squared deviation of the baseline, respectively [51,52]. The detailed calculation process and a comparison of the five sensors can be found in Tables S1–S6 in the Supporting Information. The results demonstrate that the theoretical LOD of the In₂O₃−3 °C/min−500 °C gas sensor is approximately 0.0392 ppm. This low detection capability makes it particularly suitable for applications requiring high-precision TMA detection, such as fish freshness monitoring systems.

Figure 8a illustrates the repeatability of the five sensors through five cycles of response to 100 ppm TMA at their optimal operating conditions. The response/recovery curves of the sensors show almost no variation over the five cycles, indicating that the five sensors possess an excellent repeatability. Complementing the repeatability tests, a long-term stability test is conducted on the In ₂O₃−3 °C/min−500 °C gas sensor. As shown in Figure S3, the sensor is tested against 100 ppm TMA at its optimal operating temperature for 32 days. The results indicate that the sensor’s response value fluctuates within a range of no more than 2.85%, demonstrating that the In₂O₃−3 °C/min−500 °C gas sensor exhibits an excellent long-term stability. Selectivity is a decisive factor in determining whether a gas sensor can distinguish target analytes from complex environmental mixtures. Five different VOC gases (benzene, acetone, xylene, methanol, and ethanol) are selected as interfering gases to test the selectivity of the sensors. As shown in Figure 8b, the In₂O₃−3 °C/min−500 °C gas sensor exhibits the highest response to TMA, demonstrating significant selectivity. Therefore, the In₂O₃−3 °C/min−500 °C gas sensor has a good repeatability and selectivity, suggesting that it has significant potential for practical applications.

In addition, changes in humidity can have an impact on sensor performance. The dynamic resistance curve of the In₂O₃−3 °C/min−500 °C sensor under humidity is shown in Figure S4a–d. The sensor’s initial resistance in air (R_a) and response to TMA decrease as humidity increases (Figure S4e). Water molecules are physically adsorbed on the surface of the sensing material and react with oxygen ions on the surface of the material, reducing the thickness of the depletion layer and competing with the TMA for the adsorption site, resulting in a decrease in the resistance and response of the sensor.

In summary, the In₂O₃−3 °C/min−500 °C gas sensor exhibits an optimal performance in terms of sensing response, response/recovery characteristics, LOD, selectivity, and repeatability. As shown in

Table 3. A comparison of the performance of the In₂O₃−3 °C/min−500 °C-based sensor in this work with other TMA gas sensors.

Mater.	Temp. (°C)	Conc. (ppm)	Res.	τres./τrec. (s)	Ref.
Co2O3/In2O3	200	10	11.67	25/68	[53]
In2O3@In2S3	100	10	5.5	28/48	[54]
α-Fe2O3	250	20	7.1	100/200	[33] *
Au−WO3	268	100	42.56	1/323	[34]
TiO2−NiFe2O4	307	10	11.2	50/45	[55]
NiO	260	100	13.51	24/39	[56] *
NiMoO4/MoO3	200	10	62.79	13/78	[57] *
Au/ZnO	250	10	52.6	12/127	[58] *
In2O3−NiO	200	10	20.51	39/43	[59]
3 %Bi2O3−In2O3 NFs	250	100	48.6	320/58	[60]
In2O3	240	100	47.0	6/37	This work

Mater.: materials; Temp.: operating temperature; Conc.: gas concentration; Res.: response; τ_res./τ_rec.: response/recovery time; Ref.: references; *: Literature related to freshness monitoring of seafood (fish, shrimp, etc.).

, compared to other reported TMA gas sensors, the In₂O₃−3 °C/min−500 °C gas sensor demonstrates a higher response and faster response/recovery speed, indicating a greater potential for application in assessing fish freshness.

To assess the practical application potential of the In₂O₃−500 °C−3 °C/min nanowire-based gas sensor, response measurements are conducted on L. crocea (approximately 500 g) stored in hermetically sealed glass containers under different temperature conditions (5 °C refrigeration and 25 °C ambient storage). The results show that for L. crocea stored at both temperatures, the sensor response values show an increasing trend with storage duration, which corresponds well with the fish spoilage progression. It has been reported that the gas component released during the spoilage process of L. crocea is TMA molecules [3]. A TMA concentration exceeding 10 ppm is defined as the threshold for fish spoilage. Based on the concentration–response curve fitted in Figure 7f, the In₂O₃−500 °C−3 °C/min nanowire-based gas sensor exhibits the following concentration–response function: y = 0.84362x + 2.00512. Thus, at the TMA concentration of 10 ppm, the corresponding sensor response is calculated as 10.44. Figure 9a shows that when stored at a room temperature of 25 °C for 3 days, the response value of the sensor is much greater than 10.44, reaching 58.42. Combined with the appearance and pungent smell of the L. crocea, this indicates that the L. crocea has undergone significant spoilage. Figure 9b shows that when stored at 5 °C for 10 days, the response value of the sensor basically reaches 10.44. Based on the corresponding images, it can be seen that there is no obvious spoilage on the surface of the L. crocea at this time, but a fishy smell appears, indicating that the L. crocea is gradually deteriorating. When stored for 16 days, the response value of the sensor reaches as high as 151.58. According to the corresponding image, it can be seen that the L. crocea shows serious spoilage at this time. From the above experiments, it can be seen that gas sensors based on the In₂O₃−500 °C−3 °C/min gas sensor have good application prospects in detecting fish freshness.

3.3. Gas-Sensing Mechanism

In₂O₃ is an n-type MOS material, and its mainstream gas-sensing mechanism is the electron depletion theory. This theory suggests that during the adsorption and desorption of gas molecules, the chemically adsorbed oxygen ions on the surface of the sensing material react with the gas molecules, causing a change in resistance [61,62]. When the sensor is exposed to air, oxygen molecules adsorb onto the material’s surface and capture electrons from the conduction band, forming different reactive oxygen species (${\mathrm{O}}_2^{-}$, ${\mathrm{O}}^{-}$, and ${\mathrm{O}}^{2-}$) in different temperature ranges. The optimal operating temperature for the In₂O₃−3 °C/min−500 °C nanowires in this work is 240 °C, hence, the reaction is predominantly governed by Equation (4), as follows:

$${\mathrm{O}}_2\left(\mathrm{gas}\right)\to {\mathrm{O}}_2\left(\mathrm{ads}\right)$$

(2)

$${\mathrm{O}}_2\left(\mathrm{ads}\right)+{\mathrm{e}}^{-}\to {\mathrm{O}}_2^{-}\left(\mathrm{ads}\right) (\mathrm{T}<150 °\mathrm{C})$$

(3)

$${\mathrm{O}}_2^{-}\left(\mathrm{ads}\right)+{\mathrm{e}}^{-}\to 2{\mathrm{O}}^{-}\left(\mathrm{ads}\right) (150 °\mathrm{C}<\mathrm{T}<400 °\mathrm{C})$$

(4)

$${\mathrm{O}}^{-}\left(\mathrm{ads}\right)+{\mathrm{e}}^{-}\to {\mathrm{O}}^{2-}\left(\mathrm{ads}\right) (\mathrm{T}>400 °\mathrm{C})$$

(5)

At this point, the process will lead to electron depletion in the space charge region and the formation of an electron depletion layer on the material’s surface, resulting in an increase in the material’s resistance. When the sensing material is exposed to a reducing gas (TMA) at its operating temperature, TMA molecules will undergo a redox reaction with the oxygen ions on the material, as shown in Equation (6). During this process, electrons will be released back into the conduction band of In₂O₃, which reduces the width of the electron depletion layer and lowers the resistance. The process described above is illustrated in Figure 10 a–c.

$$2{\left({\mathrm{CH}}_3\right)}_3\mathrm{N}+ 25{\mathrm{O}}^{-}\to 9{\mathrm{H}}_2\mathrm{O}+6\mathrm{C}{\mathrm{O}}_{2 }+2\mathrm{N}{\mathrm{O}}_{2 }+25{\mathrm{e}}^{-}$$

(6)

In₂O₃-based materials are widely used for NO₂ detection, but low-temperature or even room-temperature detection is the mainstream. In addition, when the sensor is exposed to a TMA atmosphere for detection, only a very small proportion of TMA molecules actually participate in the reaction. Therefore, even if the product of TMA reaction contains NO₂, the actual concentration of NO₂ in the entire test atmosphere is extremely low, expected to be in the ppb range. Nevertheless, 100 ppm of NO₂, which is much higher than its actual concentration, is still chosen as the interfering gas for selectivity tests. The results in Figure S5 show that the response of the In₂O₃−3 °C/min−500 °C gas sensor to 100 ppm NO₂ gas at 240 °C is only 6.63, still much lower than the response to TMA. Therefore, NO₂ as a product of TMA does not interfere with selectivity.

The high response of the In₂O₃−3 °C/min−500 °C nanowire-based sensor in this work can also be attributed to the following factors. Firstly, the one-dimensional structure of In₂O₃ nanowires is critical. This structure provides a high specific surface area, which allows TMA molecules greater opportunities to fully contact and react with chemisorbed oxygen on the exposed particle surfaces compared to other morphologies. Secondly, XPS analysis reveals an elevated concentration of O_V and O_C in the In₂O₃−3 °C/min−500 °C nanowires, indicating more active oxygen species and adsorption sites participating in the reaction. Additionally, SEM images demonstrate that the In₂O₃−3 °C/min−500 °C nanowires exhibit a rough surface and localized voids, which provide additional adsorption points and facilitate the rapid diffusion of TMA molecules.

For MOS-type sensors, the smaller grain size provides more active sites for gas molecules to interact with, enhancing the sensor’s gas-sensing capability [63], as confirmed by XRD analysis. Notably, both the In₂O₃−3 °C/min−500 °C and In₂O₃−3 °C/min−400 °C nanowires exhibit smaller grain sizes, contributing to their high responses to TMA. However, the In₂O₃−3 °C/min−400 °C nanowire does not achieve the highest response, suggesting that grain size is not the sole determining factor. Beyond size, the crystallinity of the nanowires also synergistically affects sensing performance [64,65]. Poorly crystallized oxide nanofibers possess more structural defects, leading to a higher concentration of charge carriers. This results in smaller resistance variations during the electron anchoring and release processes caused by gas interactions, thereby diminishing sensor response. In contrast, well-crystallized metal oxide nanofibers have fewer charge carriers, enabling significant resistance changes during gas adsorption/desorption and, thus, a superior sensor response. Comparing the crystallinity of the two samples, the In₂O₃−3 °C/min−500 °C nanowires exhibit a significantly higher crystallinity. Consequently, under the dual influence of optimal grain size and enhanced crystallinity, the In₂O₃−3 °C/min−500 °C nanowire-based sensor demonstrates the highest response.

4. Conclusions

In summary, this work employed a straightforward electrospinning method to synthesize one-dimensional In₂O₃ nanowires. By varying the heating rates and annealing temperatures during high-temperature calcination, five distinct In₂O₃ samples were obtained. Gas-sensing tests demonstrated that the In₂O₃−3 °C/min−500 °C gas sensor outperformed other samples, exhibiting a higher response (47.0), shorter response time (6 s), and lower LOD (0.0392 ppm), as well as an excellent reproducibility and long-term stability. Through comprehensive characterization, the enhanced gas-sensing performance was attributed to the smaller grain size and superior crystallinity of the In₂O₃−3 °C/min−500 °C nanowires. These microstructural features facilitated the adsorption and desorption of TMA gas, thereby improving the sensor performance. This study provides a novel approach for fabricating gas sensors tailored for TMA detection and offers practical insights for monitoring fish freshness. However, limitations remain, such as the sensor’s poor selectivity and susceptibility to interference from other VOCs. Addressing these challenges will be critical in future research.